CN117269701B - High-voltage switch cabinet partial discharge positioning method based on artificial intelligence - Google Patents
High-voltage switch cabinet partial discharge positioning method based on artificial intelligence Download PDFInfo
- Publication number
- CN117269701B CN117269701B CN202311553486.7A CN202311553486A CN117269701B CN 117269701 B CN117269701 B CN 117269701B CN 202311553486 A CN202311553486 A CN 202311553486A CN 117269701 B CN117269701 B CN 117269701B
- Authority
- CN
- China
- Prior art keywords
- component
- signal
- connotation
- frequency
- partial discharge
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 15
- 238000001228 spectrum Methods 0.000 claims abstract description 35
- 230000003595 spectral effect Effects 0.000 claims abstract description 20
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims description 13
- 238000009792 diffusion process Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 7
- 238000009825 accumulation Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 abstract description 12
- 230000008569 process Effects 0.000 description 15
- 238000002604 ultrasonography Methods 0.000 description 8
- 230000001052 transient effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 108091034117 Oligonucleotide Proteins 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000010891 electric arc Methods 0.000 description 1
- 230000005672 electromagnetic field Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 239000012774 insulation material Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1209—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing using acoustic measurements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Algebra (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Acoustics & Sound (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention relates to the technical field of partial discharge detection, and provides a high-voltage switch cabinet partial discharge positioning method based on artificial intelligence, which comprises the following steps: acquiring ultrasonic signals of the vertex angle; determining the energy density of each peak frequency based on the energy distribution characteristics in the power spectral density estimation result of the connotation modal component; determining a partial discharge signal component frequency interval according to the energy density of all peak frequencies in the power spectrum density estimation result; determining a real signal-to-noise ratio according to the frequency interval of the partial discharge signal component of the connotation modal component; determining a true prior signal-to-noise ratio based on the true signal-to-noise ratios of all the connotation modal components contained in each ultrasonic signal; obtaining an effective ultrasonic signal based on a real priori signal-to-noise ratio; and determining the position information of partial discharge in the high-voltage switch cabinet according to the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet. The invention evaluates the real priori signal-to-noise ratio of the ultrasonic signal by using the signal decomposition algorithm, and improves the accuracy of the partial discharge positioning result.
Description
Technical Field
The invention relates to the technical field of partial discharge detection, in particular to a high-voltage switch cabinet partial discharge positioning method based on artificial intelligence.
Background
A high-voltage switchgear is a device for controlling and protecting high-voltage power equipment. It is typically composed of switches, circuit breakers, fuses, disconnectors, etc. for controlling current, voltage and power in electrical power systems. Its main functions include: opening and closing circuits, protecting electrical equipment, isolating circuits, and measuring and monitoring electrical parameters. The high-voltage switch cabinet is widely applied to the fields of power systems, industrial and mining enterprises, buildings, traffic and the like, and is an important component of power equipment.
The high-voltage switch cabinet partial discharge refers to the phenomenon that an electric shock occurs in a certain partial area in an insulation system, and is usually caused by defects or damages of insulation materials, so that arc discharge can be possibly initiated, high-temperature and high-energy arc is generated, serious damages are caused to the switch cabinet and surrounding equipment, meanwhile, the partial discharge also can cause energy loss in the power system, and the efficiency of the system is reduced. The common high-voltage switch cabinet partial discharge detection method comprises a transient voltage detection method, an ultrasonic detection method and the like, wherein the position of a discharge point is judged through the transient voltage to ground generated during partial discharge during transient voltage detection, the positioning accuracy of the transient voltage detection method depends on the acquisition effect of the transient voltage, and a high requirement is placed on an acquisition instrument; when the ultrasonic detection method is used for carrying out partial discharge positioning, noise in an application scene can influence a received ultrasonic signal, so that the final positioning effect is influenced.
Disclosure of Invention
The invention provides a high-voltage switch cabinet partial discharge positioning method based on artificial intelligence, which solves the problem of positioning accuracy errors caused by positioning partial discharge positions in ultrasonic detection by interference noise, and adopts the following specific technical scheme:
the invention relates to a high-voltage switch cabinet partial discharge positioning method based on artificial intelligence, which comprises the following steps:
acquiring ultrasonic signals of each vertex angle of a high-voltage switch cabinet;
determining the energy density of each peak frequency in the power spectral density estimation result based on the energy distribution characteristics in the power spectral density estimation result of each connotation modal component; determining a partial discharge signal component frequency interval of each connotation modal component according to the energy density of all peak frequencies in the power spectrum density estimation result;
determining the real signal-to-noise ratio of each connotation modal component according to the frequency interval of the partial discharge signal component of each connotation modal component; determining a true prior signal-to-noise ratio of the ultrasonic signal of each vertex angle based on the true signal-to-noise ratios of all connotation modal components contained in the ultrasonic signal of each vertex angle;
obtaining effective ultrasonic signals of each vertex angle based on the real priori signal-to-noise ratio by adopting a wiener filtering algorithm; and determining the position information of partial discharge in the high-voltage switch cabinet according to the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet.
Preferably, the method for determining the energy density of each peak frequency in the power spectrum density estimation result based on the energy distribution characteristics in the power spectrum density estimation result of each connotation modal component comprises the following steps:
decomposing the ultrasonic signals of each vertex angle into a preset number of connotation modal components by using a signal decomposition algorithm; taking each connotation modal component as the input of a power spectral density estimation algorithm to obtain a power spectral density estimation result of each connotation modal component;
taking the frequency corresponding to each maximum point in the power spectrum density estimation result as a peak frequency, and determining a frequency window of each peak frequency in the power spectrum density estimation result according to the local density of each peak frequency in the power spectrum density estimation result;
taking the average value of the power spectrum density value of each peak frequency on the frequency window of each peak frequency as the energy density of each peak frequency.
Preferably, the method for determining the frequency window of each peak frequency in the power spectrum density estimation result according to the local density of each peak frequency in the power spectrum density estimation result comprises the following steps:
taking the frequency of each data point in the power spectrum density estimation result as an abscissa, taking a two-dimensional coordinate system constructed by taking the power spectrum density estimation value of each data point as an ordinate as a characteristic coordinate system, and obtaining the mapping result of all data points in the power spectrum density estimation result in the characteristic coordinate system;
obtaining a cluster where each data point is located based on the mapping result by adopting a density peak clustering algorithm; and determining a frequency window of each peak frequency based on the intra-class distance of the cluster in which the corresponding data point of each peak frequency is located.
Preferably, the method for determining the frequency window of each peak frequency based on the intra-class distance of the cluster where the corresponding data point of each peak frequency is located is as follows:
taking any peak frequency corresponding data point as a target point, acquiring the mean value of Euclidean distance between each target point and the rest of data in the cluster where the target point is located as a numerator, taking the maximum value of Euclidean distance between each target point and the rest of data in the cluster where the target point is located as a denominator, and taking the ratio of the numerator to the denominator as a frequency diffusion ratio;
taking the difference value of the preset parameter and the frequency diffusion ratio as a first proportionality coefficient, and taking the sum of the preset parameter and the frequency diffusion ratio as a second proportionality coefficient;
taking the rounded result of the product of each peak frequency and the first proportionality coefficient as a frequency lower limit, taking the rounded result of the product of each peak frequency and the second proportionality coefficient as a frequency upper limit, and taking a frequency interval determined by the frequency lower limit and the frequency upper limit as a frequency window of each peak frequency.
Preferably, the method for determining the frequency interval of the partial discharge signal component of each content modal component according to the energy densities of all peak frequencies in the power spectrum density estimation result includes:
and respectively acquiring peak frequencies corresponding to the maximum values in all the energy densities in the power spectrum density estimation results of each content modal component as target frequencies, and taking a frequency window corresponding to the target frequencies as a partial discharge signal component frequency interval of each content modal component.
Preferably, the method for determining the true signal-to-noise ratio of each connotation modal component according to the frequency interval of the partial discharge signal component of each connotation modal component comprises the following steps:
taking each connotation modal component with DC direct current components removed and peak frequencies in the partial discharge signal component frequency interval removed as an effective estimation component; taking the power of the effective estimated component on the ultrasonic wave band corresponding to the maximum value of the peak frequency and the average power of the signal serving as each connotation mode component;
taking each effective estimated component after removing the preset subharmonic wave band as a noise estimated component; taking the median value of the power corresponding to all frequencies in each noise estimation component as a first estimation value; taking the first estimated value as the power value of all the removal frequencies in each noise estimated component, and taking the accumulated sum of the power corresponding to all the frequencies in each noise estimated component and the power value of all the removal frequencies as the noise average power of each connotation mode component;
and obtaining the real signal-to-noise ratio of each connotation modal component based on the signal average power and the noise average power of each connotation modal component.
Preferably, the method for determining the true prior signal-to-noise ratio of the ultrasonic signal of each vertex angle based on the true signal-to-noise ratios of all the connotation modal components included in the ultrasonic signal of each vertex angle comprises the following steps:
determining component weights of each connotation mode component based on the partial discharge signal component frequency interval of each connotation mode component;
taking the product of the component weight of each connotation modal component and the real signal to noise ratio of each connotation modal component as a first accumulation factor; and taking the accumulation result of the first accumulation factor on all connotation modal components contained in the ultrasonic signal of each vertex angle as the actual prior signal-to-noise ratio of the ultrasonic signal of each vertex angle.
Preferably, the method for determining the component weight of each connotation mode component based on the frequency interval of the partial discharge signal component of each connotation mode component comprises the following steps:
taking the proportion of the partial discharge signal component frequency interval of the first connotation modal component contained in the ultrasonic signal of each vertex angle to the frequency interval of the first connotation modal component as the component weight of the first connotation modal component;
determining a weight factor of each connotation modal component based on an order value of each connotation modal component contained in the ultrasonic signal of each vertex angle;
and taking the product of the component weight of the first connotation modal component and the weight factor of each connotation modal component as the component weight of each connotation modal component.
Preferably, the method for obtaining the effective ultrasonic signal of each vertex angle based on the real priori signal-to-noise ratio by adopting a wiener filtering algorithm comprises the following steps:
the ultrasonic signals of each vertex angle are used as input, the real priori signal-to-noise ratio of the ultrasonic signals of each vertex angle is used as the signal-to-noise ratio parameter in the wiener filtering denoising algorithm, and the output of the wiener filtering denoising algorithm is used as the effective ultrasonic signals of each vertex angle.
Preferably, the method for determining the position information of the partial discharge in the high-voltage switch cabinet according to the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet comprises the following steps:
and respectively acquiring the receiving time of the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet, and acquiring the coordinate information of the partial discharge position in the high-voltage switch cabinet based on the time difference between the receiving time and the amplitude of the effective ultrasonic signals.
The beneficial effects of the invention are as follows: according to the method, the difference between the noise signal and the partial discharge signal in the ultrasonic signal of each vertex angle is analyzed, and the partial discharge signal component frequency interval corresponding to the partial discharge signal in each connotation mode component is determined based on the power spectral density estimation result of the connotation mode component; constructing component weights according to different information amounts carried by each connotation modal component in the signal decomposition process, and determining a real priori signal-to-noise ratio based on the component weights and the real signal-to-noise ratio of the connotation modal components; the method has the beneficial effects that the time difference of effective ultrasonic signal propagation between different ultrasonic sensors is obtained by removing noise signals in the ultrasonic signals based on the real priori signal-to-noise ratio, so that the positioning accuracy of partial discharge in the high-voltage switch cabinet is improved; and secondly, the position information of partial discharge in a plurality of high-voltage switch cabinets is obtained through a repeated triangular positioning method, so that errors caused by one-time positioning are avoided, and the positioning accuracy of the partial discharge coordinates in the high-voltage switch cabinets is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a schematic flow chart of a method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to an embodiment of the invention;
FIG. 2 is a schematic diagram of a power spectral density estimation result of an connotation mode component according to an embodiment of the present invention;
fig. 3 is a flowchart of an implementation of a method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a flowchart of a method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to an embodiment of the invention is shown, and the method includes the following steps:
and S001, acquiring ultrasonic signals of each vertex angle of the high-voltage switch cabinet.
In the high-voltage switch cabinet, the partial discharge process is an instantaneous burst process of energy, electric breakdown occurs in an air gap, electric energy is converted into heat energy at one instant, gas in a discharge center expands under the action of the heat energy and propagates outwards through sound waves, and therefore the invention considers that the ultrasonic signal in the high-voltage switch cabinet is utilized to determine the position of the partial discharge of the high-voltage switch cabinet.
Eight vertex angles in the high-voltage switch cabinet are respectively provided with an ultrasonic sensorThe device is used for collecting ultrasonic signals in the high-voltage switch cabinet and respectively marking the collected ultrasonic signals as:where N represents the number of ultrasonic sensors placed at the apex angle, and the value of N in the present invention is 8.
So far, the ultrasonic signals of each vertex angle of the high-voltage switch cabinet are obtained and used for subsequently determining the positioning information of partial discharge in the high-voltage switch cabinet.
Step S002, determining the energy density of each peak frequency in the power spectrum density estimation result based on the energy distribution characteristics in the power spectrum density estimation result of each connotation modal component, and determining the frequency interval of the partial discharge signal component of each connotation modal component according to the energy densities of all peak frequencies in the power spectrum density estimation result.
When partial discharge occurs in the high-voltage switch cabinet, the ultrasonic signals collected by each vertex angle contain noise besides the partial discharge signals, wherein the noise comprises environmental noise and electronic noise, and the environmental noise refers to noise generated by other equipment in the surrounding environment in the process of collecting the ultrasonic signals, such as high-frequency vibration noise caused by a cooling fan and other mechanical equipment; the electronic noise refers to noise of the ultrasonic sensor itself, and may be caused by electronic components or may be caused by current and electromagnetic field changes. When the partial discharge positioning is performed based on a plurality of ultrasonic signals, the positioning accuracy is affected by the presence of the above-mentioned various noises, so that a noise reduction process is required before the partial discharge signal is extracted from the ultrasonic signals, so as to eliminate the influence of the noise on the partial discharge signal.
The ultrasonic signal of any one vertex angle is subjected to signal decomposition processing, for example, the ultrasonic signal of the 2 nd vertex angleAs input to the empirical mode decomposition EMD (Empirical Mode Decomposition) algorithm, the EMD algorithm is used to apply the ultrasound signal +.>Decomposing into A content modal components and a residual wave, and adding +.>A>The content mode component is marked as->The EMD algorithm is a well-known technique, and the specific process is not described in detail. Because the working state of the equipment in the high-voltage switch cabinet is generally in a relatively stable state, the partial discharge phenomenon is randomly generated, and therefore, the noise signal in the ultrasonic signal is more stable compared with the partial discharge signal, and the noise signal in the connotation mode component generally has a larger spectrum loan.
Further, power spectrum density estimation is carried out on any one connotation modal component, and the energy density of each peak frequency in each connotation modal component is obtained according to the power spectrum density estimation result. Specifically, with the a-th connotation modal componentFor example, the content modality component +.>As input of the Berger Burg algorithm, the algorithm output is the connotational modal component +.>As shown in fig. 2. It should be noted that, in the present invention, the number of autoregressive models in the Burg algorithm is determined by using the red pool information criterion AIC, where the red pool information criterion AIC and the Burg algorithm are known techniques, and the specific process is not repeated.
Second, the content modal componentFrequency of each data point in the power spectral density estimation result of (2) asOn the abscissa, a two-dimensional coordinate system constructed by taking the power spectrum density estimated value of each data point as the ordinate is taken as a characteristic coordinate system, and the content modal component +.>Mapping results of all data points in the characteristic coordinate system in the power spectrum density estimation results. Further, taking all data points in the mapping result as input of a clustering algorithm, and acquiring the clustering result of the mapping result by adopting a density peak clustering algorithm. In the clustering process, the data points belonging to noise signals are divided into the same cluster, and the more concentrated the noise, the more likely the data points in the frequency interval become a cluster center point; similarly, the data points belonging to the partial discharge signals are divided into the same cluster, and the local density of the data points in the cluster corresponding to the partial discharge signals is larger due to the randomness of the partial discharge signals, so that the density peak clustering algorithm is a known technology, and the specific process is not repeated.
Specifically, the content modal component is acquired using the peak detection algorithm PDA (Peak detection algorithm)All maxima points in the power spectral density estimation result of (2) and incorporating the modal component +.>Frequency corresponding to each maximum point in the power spectral density estimation result of (a) is taken as an connotation mode component +.>Is to include the modal component +.>The frequency corresponding to the ith extreme point of (2) is denoted as peak frequency +.>. The peak detection algorithm PDA is a well-known technique, and the specific detection process is not described in detail. According to the corresponding number of each peak frequencyDetermining a frequency window of each peak frequency by using a cluster where the data points are located, and marking the corresponding data points as +.>Data point +.>The cluster in which this is located is marked +.>Calculating peak frequency +.>Corresponding frequency window->，/>Wherein->、/>The lower frequency limit and the upper frequency limit of the frequency window are respectively. The specific calculation formula is as follows: />
In the method, in the process of the invention,is peak frequency +.>Frequency diffusion ratio of>Is data point +.>Cluster->Mean value of Euclidean distance between all data points in the inner, < >>Is data point +.>Cluster->A maximum value of Euclidean distance between all data points; />、/>Peak frequency +.>Frequency lower limit and frequency upper limit of frequency window>Is a rounding function.
Wherein, clusterThe greater the local density of the internal data points, +.>The smaller the value of (2), the data point +.>Cluster->Between all data points inMean value of Euclidean distance>And maximum value->The closer the (the)>The larger the first scale factorThe smaller the value of (2), the corresponding lower frequency limit +.>The smaller the value of (2); second scaling factor->The larger the value of (2), the corresponding upper frequency limit +.>The larger the value of (2), the larger the frequency window in which the peak frequency is located, the more similar the peak frequency is to the power spectral density estimation value of the adjacent frequency, i.e.)>The larger the range of (2) is, the inclusion modality component +.>The more likely it is that the partial discharge signal with a higher partial density is in the power spectral density estimation result.
According to the steps, the content modal components are respectively obtainedA frequency window corresponding to each peak frequency. Calculating the energy density of each peak frequency based on the size of the frequency window in which the peak frequency is located, and calculating the peak frequency +.>Energy density>：
In the method, in the process of the invention,is peak frequency +.>Energy density of>、/>Respectively the frequency intervalLower frequency limit, upper frequency limit, +.>Is a frequency value in the frequency interval, and it is to be noted that: />Meaning that the sum of the power spectral density estimates for all frequencies within the frequency interval is calculated, since the power spectral density estimate is continuous, rather than discrete.
According to the steps, the content modal components are respectively obtainedObtaining the connotation mode component +.>Maximum value in all corresponding energy densities, and taking a frequency window in which the peak frequency corresponding to the maximum value of the energy density is located as an connotation mode component +.>Is included in the partial discharge signal component frequency interval.
The partial discharge signal component frequency interval of each connotation mode component is used for the subsequent denoising treatment of the ultrasonic signal.
Step S003, determining the real signal-to-noise ratio of each connotation modal component according to the frequency interval of the partial discharge signal component of each connotation modal component; the true prior signal-to-noise ratio of the ultrasound signal for each vertex angle is determined based on the true signal-to-noise ratios of all the connotatory modal components contained by the ultrasound signal for each vertex angle.
For the ultrasonic signals of each vertex angle, the noise signals contained in the ultrasonic signals are unknown, so the invention obtains the signal average power and the noise average power in each connotation mode component in each ultrasonic signal by means of parameter estimation. For example, in terms of content modal componentsFor example, the content modal component is first removed by the direct differentiation method +.>The DC component in (2) is removed, then all peak frequencies contained in the partial discharge signal component frequency interval of the connotation mode component are removed on the basis of removing the DC component, and the result obtained through the steps is taken as the connotation mode component->Is>. Since each ultrasonic signal includes not only partial discharge signal, noise signal but also fundamental harmonic signal, in order to obtain more practical estimation result, the component is estimated from the effectiveRemove->A subharmonic band, wherein the result of removing the harmonic band is taken as an connotation modal componentNoise estimation component->In the present invention, < >>、/>The sizes of the high-voltage switch cabinet are respectively taken as an experience value of 2 and an experience value of 6, and an implementer can select proper +.>、/>. Based on the effective estimated component->Noise estimation component->Acquiring content modality component->Is>：
In the method, in the process of the invention,is an connotation mode component->Signal average power,/,> 、/>the effective estimated components +.>The maximum value of the middle peak frequency corresponds to the upper limit and the lower limit of the frequency interval of the ultrasonic wave band, and the maximum value of the middle peak frequency is +.>Is the frequency interval of the ultrasonic wave band +.>A power value of any one of the frequencies;
is an connotation mode component->Noise average power, < >>、/>Respectively noise estimation componentsMaximum, minimum of medium frequency, +.>Is the power value of any non-removal frequency in the frequency interval; j is the noise estimation component +.>Frequency of j-th removed, +.>Is the power value of the j-th removed frequency, is->Is equal to the noise estimate component +.>The median value of the power corresponding to all non-removal frequencies in (a), n being the number of removal frequencies;
is an connotation mode component->Is>Is a logarithmic function with a base of 10.
Wherein the content modality componentThe smaller the peak frequency contained in the partial discharge signal component frequency interval, the effective estimation component +.>The inclusion modality component contained in (a)>The more frequencies in->The greater the value of (2); the more frequencies are removed from the noise estimation component, the more the power value of the removed frequencies is equal to the noise estimation component +.>The median value of all non-removed frequencies corresponding to the power, < >>The smaller the value of (2), the corresponding true signal to noise ratio +.>The greater the value of (2).
According to the steps, the real signal-to-noise ratio of each connotation modal component contained in the ultrasonic signal of each vertex angle is obtained respectively, and the real priori signal-to-noise ratio of the ultrasonic signal of each vertex angle is obtained based on the real signal-to-noise ratios of all connotation modal components contained in the ultrasonic signal of each vertex angle. Calculating ultrasonic signals of the 2 nd apex angleTrue a priori signal to noise ratio of (2)：
In the method, in the process of the invention,is an ultrasonic signal +.>Component weights comprising a first connotation modal component,/->、Is an ultrasonic signal +.>The method comprises the steps that the frequency upper limit and the frequency lower limit of a frequency interval of partial discharge signal components corresponding to a first connotation mode component are included; />、/>Respectively ultrasonic signals->The method comprises the steps of including a frequency maximum value and a frequency minimum value in a first connotation mode component;
is an ultrasonic signal +.>Component weights comprising the b-th content modal component;
is an ultrasonic signal +.>Is the true a priori signal-to-noise ratio of (A) is the ultrasound signal +.>Number of inclusion modality components, +.>Is an ultrasonic signal +.>The true signal to noise ratio of the b-th content modal component is included.
Wherein the method comprises the steps ofUltrasonic signalThe larger the range of the frequency interval of the partial discharge signal component corresponding to the first connotation mode component is, the ultrasonic signal is +>The less noise content is contained in the first content mode component, the ultrasound signalThe greater the effect of the true snr containing the first content modal component on the signal-to-noise ratio of the ultrasound signal,the greater the value of +.>The greater the value of (2); in the process of decomposing signals in an empirical mode, the lower the order of the inclusion mode components is, the frequency components contained in the inclusion mode components are gradually reduced, and the smaller the influence of the actual signal-to-noise ratio of the inclusion mode components on the signal-to-noise ratio of the ultrasonic signal is,/>The greater the value of (2), the weighting factor +.>The smaller the value of +.>The smaller the value of (2), the first accumulation factor +.>The smaller the value of (2).
Thus, the real priori signal-to-noise ratio of the ultrasonic signals of each vertex angle is obtained and is used for subsequently obtaining the effective ultrasonic signals of each vertex angle.
Step S004, obtaining effective ultrasonic signals of each vertex angle based on the real priori signal-to-noise ratio by adopting a wiener filtering algorithm; and determining the position information of partial discharge in the high-voltage switch cabinet according to the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet.
According to the steps, the real priori signal-to-noise ratio of the ultrasonic signals of each vertex angle is obtained respectively. Secondly, respectively taking the ultrasonic signals of each vertex angle as the input of a wiener filtering denoising algorithm, taking the real priori signal-to-noise ratio of the ultrasonic signals as the signal-to-noise ratio parameter in the wiener filtering denoising algorithm, outputting the effective ultrasonic signals after denoising the ultrasonic signals of each vertex angle by using the wiener filtering denoising algorithm, and obtaining the ultrasonic signalsThe corresponding effective ultrasound signal is marked +.>The wiener filtering denoising algorithm is a known technology, and the specific process is not repeated.
Further, the position information of partial discharge in the high-voltage switch cabinet is obtained according to the effective ultrasonic signals of each vertex angle on the high-voltage switch cabinet, and the specific implementation flow is shown in fig. 3. For any ultrasonic sensor at any vertex angle, acquiring the transmitting time of an ultrasonic signal sent to a high-voltage switch cabinet by using the ultrasonic sensor, and then taking any other ultrasonic sensor as a signal receiver, and acquiring the receiving time of an effective ultrasonic signal by using the signal receiver. For example, the ultrasonic sensor at the 2 nd vertex angle transmits an ultrasonic signal to the high-voltage switch cabinetAnd ultrasound signal->Is recorded as +.>The 1 st vertex angle ultrasonic sensor is used as a signal receiver, and the 1 st vertex angle ultrasonic sensor receives effective ultrasonic signals +.>Is recorded as +.>Time difference between ultrasonic sensors at 2 nd apex angle and 1 st apex angle based on ultrasonic signal +.>And the propagation speed of the ultrasonic signal obtains the spatial distance between the ultrasonic sensors of the 2 nd vertex angle and the 1 st vertex angle +.>. And repeating the flow to obtain the space distance D between any two vertex angle sensors on the high-voltage switch cabinet.
And secondly, calculating the coordinate information of partial discharge based on the space distance between two vertex angle sensors on the high-voltage switch cabinet by using a triangle positioning method. For example, coordinate information of the ultrasonic sensor at the 1 st vertex angle, the 2 nd vertex angle, and the N th vertex angle are respectively recorded as、/>、/>The coordinates of partial discharge calculated by the first triangulation method are set as +.>According to->、/>And the spatial distance between the ultrasonic sensors of the 2 nd apex angle and the 1 st apex angle +.>Can obtain a group of ternary one timeEquation:
according to the three-dimensional once equation of each group, a partial discharge coordinate can be determined, in order to improve the reliability of a positioning result, in the invention, N groups of partial discharge coordinates are continuously calculated, the size of N is taken as a checked value 6, the average value of the N groups of partial discharge coordinates is taken as a final positioning coordinate, and an operator takes necessary maintenance measures according to the final positioning coordinate to ensure the safe operation of the high-voltage switch cabinet.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (6)
1. The high-voltage switch cabinet partial discharge positioning method based on artificial intelligence is characterized by comprising the following steps of:
acquiring ultrasonic signals of each vertex angle of a high-voltage switch cabinet;
determining the energy density of each peak frequency in the power spectral density estimation result based on the energy distribution characteristics in the power spectral density estimation result of each connotation modal component; determining a partial discharge signal component frequency interval of each connotation modal component according to the energy density of all peak frequencies in the power spectrum density estimation result;
determining the real signal-to-noise ratio of each connotation modal component according to the frequency interval of the partial discharge signal component of each connotation modal component; determining a true prior signal-to-noise ratio of the ultrasonic signal of each vertex angle based on the true signal-to-noise ratios of all connotation modal components contained in the ultrasonic signal of each vertex angle;
obtaining effective ultrasonic signals of each vertex angle based on the real priori signal-to-noise ratio by adopting a wiener filtering algorithm; the method for determining the energy density of each peak frequency in the power spectrum density estimation result based on the energy distribution characteristics in the power spectrum density estimation result of each connotation modal component comprises the following steps of:
decomposing the ultrasonic signals of each vertex angle into a preset number of connotation modal components by using a signal decomposition algorithm; taking each connotation modal component as the input of a power spectral density estimation algorithm to obtain a power spectral density estimation result of each connotation modal component;
taking the frequency corresponding to each maximum point in the power spectrum density estimation result as a peak frequency, and determining a frequency window of each peak frequency in the power spectrum density estimation result according to the local density of each peak frequency in the power spectrum density estimation result; taking the average value of the power spectrum density value of each peak frequency on the frequency window of each peak frequency as the energy density of each peak frequency, wherein the method for determining the real signal-to-noise ratio of each connotation modal component according to the partial discharge signal component frequency interval of each connotation modal component comprises the following steps:
taking each connotation modal component with DC direct current components removed and peak frequencies in the partial discharge signal component frequency interval removed as an effective estimation component; taking the power of the effective estimated component on the ultrasonic wave band corresponding to the maximum value of the peak frequency and the average power of the signal serving as each connotation mode component;
taking each effective estimated component after removing the preset subharmonic wave band as a noise estimated component; taking the median value of the power corresponding to all frequencies in each noise estimation component as a first estimation value; taking the first estimated value as the power value of all the removal frequencies in each noise estimated component, and taking the accumulated sum of the power corresponding to all the frequencies in each noise estimated component and the power value of all the removal frequencies as the noise average power of each connotation mode component;
the real signal-to-noise ratio of each connotation mode component is obtained based on the signal average power and the noise average power of each connotation mode component, and the method for determining the real priori signal-to-noise ratio of the ultrasonic signal of each vertex angle based on the real signal-to-noise ratios of all connotation mode components contained in the ultrasonic signal of each vertex angle is as follows:
determining component weights of each connotation mode component based on the partial discharge signal component frequency interval of each connotation mode component; taking the product of the component weight of each connotation modal component and the real signal to noise ratio of each connotation modal component as a first accumulation factor; the method for determining the position information of partial discharge in the high-voltage switch cabinet according to the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet comprises the following steps of:
and respectively acquiring the receiving time of the effective ultrasonic signals of all the vertex angles of the high-voltage switch cabinet, and acquiring the coordinate information of the partial discharge position in the high-voltage switch cabinet based on the time difference between the receiving time and the amplitude of the effective ultrasonic signals.
2. The method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to claim 1, wherein the method for determining a frequency window of each peak frequency in the power spectral density estimation result according to the local density of each peak frequency in the power spectral density estimation result is as follows:
taking the frequency of each data point in the power spectrum density estimation result as an abscissa, taking a two-dimensional coordinate system constructed by taking the power spectrum density estimation value of each data point as an ordinate as a characteristic coordinate system, and obtaining the mapping result of all data points in the power spectrum density estimation result in the characteristic coordinate system;
obtaining a cluster where each data point is located based on the mapping result by adopting a density peak clustering algorithm; and determining a frequency window of each peak frequency based on the intra-class distance of the cluster in which the corresponding data point of each peak frequency is located.
3. The method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to claim 2, wherein the method for determining the frequency window of each peak frequency based on the intra-class distance of the cluster where the corresponding data point of each peak frequency is located is as follows:
taking any peak frequency corresponding data point as a target point, acquiring the mean value of Euclidean distance between each target point and the rest of data in the cluster where the target point is located as a numerator, taking the maximum value of Euclidean distance between each target point and the rest of data in the cluster where the target point is located as a denominator, and taking the ratio of the numerator to the denominator as a frequency diffusion ratio;
taking the difference value of the preset parameter and the frequency diffusion ratio as a first proportionality coefficient, and taking the sum of the preset parameter and the frequency diffusion ratio as a second proportionality coefficient;
taking the rounded result of the product of each peak frequency and the first proportionality coefficient as a frequency lower limit, taking the rounded result of the product of each peak frequency and the second proportionality coefficient as a frequency upper limit, and taking a frequency interval determined by the frequency lower limit and the frequency upper limit as a frequency window of each peak frequency.
4. The method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to claim 1, wherein the method for determining the frequency interval of the partial discharge signal component of each connotation modal component according to the energy density of all peak frequencies in the power spectrum density estimation result is as follows:
and respectively acquiring peak frequencies corresponding to the maximum values in all the energy densities in the power spectrum density estimation results of each content modal component as target frequencies, and taking a frequency window corresponding to the target frequencies as a partial discharge signal component frequency interval of each content modal component.
5. The method for positioning partial discharge of a high-voltage switch cabinet based on artificial intelligence according to claim 1, wherein the method for determining the component weight of each connotation mode component based on the partial discharge signal component frequency interval of each connotation mode component is as follows:
taking the proportion of the partial discharge signal component frequency interval of the first connotation modal component contained in the ultrasonic signal of each vertex angle to the frequency interval of the first connotation modal component as the component weight of the first connotation modal component;
determining a weight factor of each connotation modal component based on an order value of each connotation modal component contained in the ultrasonic signal of each vertex angle;
and taking the product of the component weight of the first connotation modal component and the weight factor of each connotation modal component as the component weight of each connotation modal component.
6. The high-voltage switch cabinet partial discharge positioning method based on artificial intelligence according to claim 1, wherein the method for obtaining the effective ultrasonic signal of each vertex angle based on the real priori signal-to-noise ratio by adopting a wiener filtering algorithm is as follows:
the ultrasonic signals of each vertex angle are used as input, the real priori signal-to-noise ratio of the ultrasonic signals of each vertex angle is used as the signal-to-noise ratio parameter in the wiener filtering denoising algorithm, and the output of the wiener filtering denoising algorithm is used as the effective ultrasonic signals of each vertex angle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311553486.7A CN117269701B (en) | 2023-11-21 | 2023-11-21 | High-voltage switch cabinet partial discharge positioning method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311553486.7A CN117269701B (en) | 2023-11-21 | 2023-11-21 | High-voltage switch cabinet partial discharge positioning method based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117269701A CN117269701A (en) | 2023-12-22 |
| CN117269701B true CN117269701B (en) | 2024-02-02 |
Family
ID=89210888
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311553486.7A Active CN117269701B (en) | 2023-11-21 | 2023-11-21 | High-voltage switch cabinet partial discharge positioning method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117269701B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5983162A (en) * | 1996-08-12 | 1999-11-09 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Computer implemented empirical mode decomposition method, apparatus and article of manufacture |
| CN102175768A (en) * | 2011-02-22 | 2011-09-07 | 哈尔滨工业大学 | Method and device for detecting defects and failures of high-speed rail based on vibration signals |
| CN110909480A (en) * | 2019-11-29 | 2020-03-24 | 华北水利水电大学 | Denoising method and denoising device for water turbine vibration signal |
| CN113702043A (en) * | 2021-08-09 | 2021-11-26 | 大连理工大学 | Planetary gear box fault diagnosis method based on POVMD and FDTW under time-varying rotation speed |
| CN114089138A (en) * | 2021-11-26 | 2022-02-25 | 平顶山天安煤业股份有限公司 | High-voltage cable partial discharge online monitoring method and system |
| CN114330067A (en) * | 2021-12-30 | 2022-04-12 | 南昌大学 | Soft foundation sluice finite element model correction method |
| CN114626193A (en) * | 2021-12-30 | 2022-06-14 | 南昌大学 | Method for reducing noise of leakage structure vibration signal based on improved variational modal decomposition |
| CN115034263A (en) * | 2022-05-31 | 2022-09-09 | 重庆大学 | Variable-scale evolution self-adaptive noise elimination method, noise elimination system and fault diagnosis system |
| CN115902528A (en) * | 2023-02-21 | 2023-04-04 | 华东交通大学 | Direct-current traction network oscillation and short-circuit fault identification method |
| CN116559643A (en) * | 2023-04-27 | 2023-08-08 | 山东电工电气集团有限公司 | GIS partial discharge fault discrimination method based on feature similarity |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1579608A4 (en) * | 2002-10-11 | 2012-09-05 | Flint Hills Scient Llc | Method, computer program, and system for intrinsic timescale decomposition, filtering, and automated analysis of signals of arbitrary origin or timescale |
| WO2018053525A1 (en) * | 2016-09-19 | 2018-03-22 | The University Of New Hampshire | Techniques for empirical mode decomposition (emd)-based noise estimation |
| CN111562358B (en) * | 2020-05-06 | 2021-03-16 | 武汉大学 | Transformer oil gas content prediction method and system based on combined model |
| CN111551992B (en) * | 2020-05-12 | 2021-02-26 | 中国科学院地质与地球物理研究所 | Rock reservoir structure characterization method and device, computer-readable storage medium and electronic equipment |
| CN112858473B (en) * | 2020-12-25 | 2022-07-12 | 杭州浙达精益机电技术股份有限公司 | Turnout switch blade damage state monitoring method based on feature fusion |
-
2023
- 2023-11-21 CN CN202311553486.7A patent/CN117269701B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5983162A (en) * | 1996-08-12 | 1999-11-09 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Computer implemented empirical mode decomposition method, apparatus and article of manufacture |
| CN102175768A (en) * | 2011-02-22 | 2011-09-07 | 哈尔滨工业大学 | Method and device for detecting defects and failures of high-speed rail based on vibration signals |
| CN110909480A (en) * | 2019-11-29 | 2020-03-24 | 华北水利水电大学 | Denoising method and denoising device for water turbine vibration signal |
| CN113702043A (en) * | 2021-08-09 | 2021-11-26 | 大连理工大学 | Planetary gear box fault diagnosis method based on POVMD and FDTW under time-varying rotation speed |
| CN114089138A (en) * | 2021-11-26 | 2022-02-25 | 平顶山天安煤业股份有限公司 | High-voltage cable partial discharge online monitoring method and system |
| CN114330067A (en) * | 2021-12-30 | 2022-04-12 | 南昌大学 | Soft foundation sluice finite element model correction method |
| CN114626193A (en) * | 2021-12-30 | 2022-06-14 | 南昌大学 | Method for reducing noise of leakage structure vibration signal based on improved variational modal decomposition |
| CN115034263A (en) * | 2022-05-31 | 2022-09-09 | 重庆大学 | Variable-scale evolution self-adaptive noise elimination method, noise elimination system and fault diagnosis system |
| CN115902528A (en) * | 2023-02-21 | 2023-04-04 | 华东交通大学 | Direct-current traction network oscillation and short-circuit fault identification method |
| CN116559643A (en) * | 2023-04-27 | 2023-08-08 | 山东电工电气集团有限公司 | GIS partial discharge fault discrimination method based on feature similarity |
Non-Patent Citations (3)
| Title |
|---|
| 基于改进CEEMD的超声检测信号自适应降噪;孙灵芳;徐曼菲;朴亨;李霞;;振动与冲击(20);全文 * |
| 基于改进经验模态分解的局部放电去噪方法;孟稳;;工业控制计算机(04);全文 * |
| 基于集合经验模态分解法的局部放电信号去噪;尚海昆;王坤;李峰;;东北电力大学学报(04);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117269701A (en) | 2023-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2541023A1 (en) | Method of detecting the position of a wave front in a signal received by a detector | |
| CN111024209B (en) | Line spectrum detection method suitable for vector hydrophone | |
| CN109142896B (en) | Lightning early warning method based on three-dimensional atmospheric electric field and MEMD | |
| KR101550689B1 (en) | A arc or corona detection system for a distributing board with the acoustic emission sensor and noise removal function | |
| CN102736070A (en) | Method for removing noises of wind profile radar based on Hilbert-Huang transform | |
| CN108646248B (en) | Passive acoustic speed and distance measuring method for low-speed moving sound source | |
| CN109991520A (en) | A kind of cable oscillation wave partial discharge detecting system velocity of wave New calculating method | |
| CN110398722A (en) | Extension target echo detection method based on the limited spectrum of random matrix | |
| CN114113943A (en) | Transformer partial discharge detection system, method and equipment based on current and ultrasonic signals | |
| CN117269701B (en) | High-voltage switch cabinet partial discharge positioning method based on artificial intelligence | |
| CN108920760A (en) | Antenna high-altitude electromagnetic pulse coupling response performance accounting method based on polynomial chaos expression | |
| CN113406441B (en) | Flexible direct-current power grid fault location method based on clustering and iterative algorithm | |
| CN116207708B (en) | Transformer fault defense method and device | |
| CN107040269B (en) | Pole based on variance median filtering/ultralow frequency channel atmospheric noise suppressing method | |
| CN115656806A (en) | Isolator monitoring method related to surface area of object | |
| CN115313329A (en) | Frequency band range selection method for rapid traveling wave protection and related device | |
| CN113608190B (en) | Sea surface target detection method and system based on three characteristics of singular space | |
| CN112487714B (en) | Method for generating cable shaft fire state identification decision tree model | |
| CN114745027A (en) | Power line communication impulse noise identification method and system, and storage medium | |
| CN109344814A (en) | A kind of GIS machine performance signal characteristic extracting methods | |
| Jiang et al. | De-noising for partial discharge signals of power transformers using wavelet threshold method based on ant colony optimization | |
| Ma et al. | Location of Breakdown Discharge of GIS Using Acoustic Signals | |
| Zmarzly et al. | High voltage power transformer diagnostics using vibroacoustic method | |
| Jung et al. | Discrimination of insulation defects in a gas insulated switchgear (GIS) by use of a neural network based on a chaos analysis of partial discharge (CAPD) | |
| CN117741356B (en) | Lightning arrester valve plate aging evaluation method, medium and system under complex working conditions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |